Heuristic search is a fundamental component of Artificial Intelligence applications. Because search routines are frequently also a computational bottleneck, numerous methods have been explored to increase the efficiency of search. Recently, researchers have begun investigating methods of using parallel MIMD and SIMD hardware to speed up the search process. In this paper, we present a massively-parallel SIMD approach to search named MIDA* search. The components of MIDA* include a very fast distribution algorithm which biases the search to one side of the tree, and an incrementally-deepening depthfirst search of all the processors in parallel. We show the results of applying MIDA* to instances of the Fifteen Puzzle problem and to the robot arm motion planning problem. Results reveal an efficiency of 74% and a speedup of 8553 and 492 over serial and 16-processor MIMD algorithms, respectively, when finding a solution to the Fifteen Puzzle problem that is close to optimal.
Single- and dual-arm motion planning with heuristic search